holisticai.bias.mitigation.MCMF#

class holisticai.bias.mitigation.MCMF(*args, **kargs)[source]#

Minimal Cluster Modification for Fairnes (MCMF) [1] is focused on the minimal change it so that the clustering is still of good quality and fairer.

Parameters

metricstr: Measure function used in the objective function. The metrics available are: [“constant”, “L1”, “L2”]
solverstr: Algorithm name used to solve the standard form problem. Solver supported must depend of your scipy package version. for scipy 1.9.0 the solvers available are: [“highs”, “highs-ds”, “highs-ipm”]
group_modestr: Set what groups will be fitted: [‘a’, ‘b’, ‘ab’]
verboseint: If > 0 , then print logs.

Examples

>>> from holisticai.bias.mitigation import MCMF
>>> mitigator = MCMF(**params)
>>> mitigator.fit_transform(X, y_pred, group_a, group_b)
>>> test_data_transformed = mitigator.transform(X, y_pred, group_a, group_b)

References

fit()[source]#

Fit the MCMF algorithm.

Returns

self

fit_transform(X: np.ndarray, y_pred: np.ndarray, group_a: np.ndarray, group_b: np.ndarray, centroids: np.ndarray) → dict[str, np.ndarray][source]#

Fit and transform the MCMF algorithm.

Description

Fit the MCMF algorithm and transform the predicted vector.

Parameters

Xmatrix-like: Input matrix (nb_examples, nb_features)
y_predarray-like: Predicted vector (nb_examples,)
group_aarray-like: Group membership vector (binary)
group_barray-like: Group membership vector (binary)
centroidsndarray: Centroids array

Returns

dict: A dictionary with new predictions

transform(X: np.ndarray, y_pred: np.ndarray, group_a: np.ndarray, group_b: np.ndarray, centroids: np.ndarray) → dict[str, np.ndarray][source]#

Modify y_pred using MCMF algorithm.

Description

Build parameters for the objetive function and call the solver to find the algorithm parameters.

Parameters

Xmatrix-like: Input matrix (nb_examples, nb_features)
y_predarray-like: Predicted vector (nb_examples,)
group_aarray-like: Group membership vector (binary)
group_barray-like: Group membership vector (binary)
centroidsndarray: Centroids array

Returns

dict: A dictionary with new predictions